I am upgrading from 9.0 to 9.1 and I completed upgrading the first host out of my 3-node cluster.
When I gave the host a reboot command, it started attempting to migrate the VMs, but it ran into this issue with every VM on that host:
The loop appears to be going continuously, cycling through all of the existing VMs. I cannot log into the host anymore via ssh since it's being rebooted, nor is a console visible anymore.
How can I stop this endless loop and allow the host to reboot?
I tried disabling all of the HA Node Affinity Rules, but that didn't stop it. I don't know how to stop this process on a node I can no longer access.
When I gave the host a reboot command, it started attempting to migrate the VMs, but it ran into this issue with every VM on that host:
Code:
task started by HA resource agent
2025-12-10 15:38:33 conntrack state migration not supported or disabled, active connections might get dropped
2025-12-10 15:38:33 starting migration of VM 105 to node 'VMHost4' (10.10.0.14)
2025-12-10 15:38:33 starting VM 105 on remote node 'VMHost4'
2025-12-10 15:38:35 [VMHost4] Unknown option: nets-host-mtu
2025-12-10 15:38:35 [VMHost4] 400 unable to parse option
2025-12-10 15:38:35 [VMHost4] qm start <vmid> [OPTIONS]
2025-12-10 15:38:35 ERROR: online migrate failure - target node VMHost4 is too old for preserving VirtIO-net MTU, please upgrade
2025-12-10 15:38:35 aborting phase 2 - cleanup resources
2025-12-10 15:38:35 migrate_cancel
2025-12-10 15:38:37 ERROR: migration finished with problems (duration 00:00:04)
TASK ERROR: migration problemsThe loop appears to be going continuously, cycling through all of the existing VMs. I cannot log into the host anymore via ssh since it's being rebooted, nor is a console visible anymore.
How can I stop this endless loop and allow the host to reboot?
I tried disabling all of the HA Node Affinity Rules, but that didn't stop it. I don't know how to stop this process on a node I can no longer access.
Last edited: